Synthesis Methodology of AW filters for RF applications based on Matrix Rotations to overcome round-off errors
Lluís Acosta, Eloi Guerrero, Patricia Silveira, Jordi Verdú, Pedro de Paco
Abstract
Filtering synthesis methodologies have become increasingly popular among AW filter designers due to their effectiveness, precision, and usefulness. Regardless of dealing with an equivalent Butterworth-Van Dyke electrical or coupling-nodal model, current synthesis tools consist of recursive methods that iteratively extract element by element the information from a global admittance function of the network. These methodologies have proven effective in finding reliable and technology-oriented seeds speedily in early design stages and providing a trustworthy tool for decision-making during the whole engineering design cycle. However, current numerical algorithms introduce roundoff errors, limiting the reachable maximum filter order. This work describes a new and completely different algorithm for the synthesis methodology that provides robustness in front of high-order filters pushing the filter degree limit beyond the 18th-order. This contribution becomes a step forward in consolidating AW-oriented EDA tools.